Autoencoder-based anomaly root cause analysis for wind turbines

作     者：Cyriana M.A.Roelofs Marc-Alexander Lutz Stefan Faulstich Stephan Vogt 

作者机构：Fraunhofer IEEKonigstor 59Kassel 34119Germany Intelligent Embedded SystemsUniversitat KasselWilhelmshoher Allee 67Kassel 34121Germany 

出 版 物：《Energy and AI》 (能源与人工智能（英文）)

年 卷 期：2021年第4卷第2期

页      面：57-65页

核心收录：

学科分类：08[工学] 081101[工学-控制理论与控制工程] 0811[工学-控制科学与工程] 081102[工学-检测技术与自动化装置] 

基　　金：German Federal Ministry for Economic Affairs and Energy, (0324128) Hessian Ministry of Higher Education , Research, Science and the Arts, (511/17.001) Hessian Ministry of Higher Education, Research, Science and the Arts 

主　　题：Anomaly detection Autoencoder Root cause analysis Predictive maintenance Wind turbine Explainable artificial intelligence 

摘      要：A popular method to detect anomalous behaviour or specific failures in wind turbine sensor data uses a specific type of neural network called an *** models have proven to be very successful in detecting such deviations,yet cannot show the underlying cause or failure *** information is necessary for the implementation of these models in the planning of maintenance *** this paper we introduce a novel method:*** use ARCANA to identify the possible root causes of anomalies detected by an *** describes the process of reconstruction as an optimisation problem that aims to remove anomalous properties from an anomaly *** reconstruction must be similar to the anomaly and thus identify only a few,but highly explanatory anomalous features,in the sense of Ockham’s *** proposed method is applied on an open data set of wind turbine sensor data,where an artificial error was added onto the wind speed sensor measurements to acquire a controlled test *** results are compared with the reconstruction errors of the autoencoder *** ARCANA method points out the wind speed sensor correctly with a significantly higher feature importance than the other features,whereas using the non-optimised reconstruction error does *** though the deviation in one specific input feature is very large,the reconstruction error of many other features is large as well,complicating the interpretation of the detected ***,we apply ARCANA to a set of offshore wind turbine *** case studies are discussed,demonstrating the technical relevance of ARCANA.